A chiplet-based gesture recognition system with modular architecture

An ultrasonic gesture recognition baseband system implemented with multiple modular dies (chiplets) is proposed in order to lower the cost by chip-level reusability. It adopts a special inter-chiplet network on active interposer, which can integrate up to 4 heterogenous chiplets. Two of them are tailorable through power management supported by the sockets on the interposer. The socket of the interposer has standard interface with 91.2 Gbps bandwidth. A CPU chiplet and a preprocessing chiplet have been designed with 28nm technology. And a modular architecture of system as well as organization method of this system are also presented in this paper. The implementation of these chiplets, the active interposer and software have been completed. The results of analysis evaluation are that it is able to recognize hand gestures with an accuracy rate around 90% and that it can support the maximum frame rate up to 6286fps, which is about 2.18 times of the existing system based on a mobile CPU.

Authors

Institutions

Publication Details

Published
2024-10-16
DOI
https://doi.org/10.1117/12.3034006
Citations
1
Primary Topic
Hand Gesture Recognition Systems
Type
article
Field-Weighted Citation Impact
0.37
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

A chiplet-based gesture recognition system with modular architecture

Yang Zheng, Xiangyu Li, Xian Tang, Yafei Liu
1 citations
Hand Gesture Recognition Systems
0.37
article

A chiplet-based gesture recognition system with modular architecture

Yang Zheng, Xiangyu Li, Xian Tang, Yafei Liu
article en
1 citations

Abstract

An ultrasonic gesture recognition baseband system implemented with multiple modular dies (chiplets) is proposed in order to lower the cost by chip-level reusability. It adopts a special inter-chiplet network on active interposer, which can integrate up to 4 heterogenous chiplets. Two of them are tailorable through power management supported by the sockets on the interposer. The socket of the interposer has standard interface with 91.2 Gbps bandwidth. A CPU chiplet and a preprocessing chiplet have been designed with 28nm technology. And a modular architecture of system as well as organization method of this system are also presented in this paper. The implementation of these chiplets, the active interposer and software have been completed. The results of analysis evaluation are that it is able to recognize hand gestures with an accuracy rate around 90% and that it can support the maximum frame rate up to 6286fps, which is about 2.18 times of the existing system based on a mobile CPU.

Tsinghua University (CN)
Openalex Percentile: Top 38%
Hand Gesture Recognition Systems
0.37
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.